StarTech
An automated-trading control room - the most architecturally complex frontend I have shipped.
StarTech is a private client engagement. The live platform is the proof; implementation details are available to walk through under NDA.
A Python trading engine executes and manages strategies automatically across crypto, forex, stocks, indices and commodities. It needed a control room: a way for operators to create requests, watch live trades, and monitor engine health in real time.
A Next.js operator dashboard sitting on top of a shared MongoDB command/state database. It surfaces everything the engine is doing - requests, live trades, positions, performance and health - across 12+ collections.
Next.js + TypeScript control room, the operator-facing layer of a two-part system
Shared MongoDB across 12+ collections, polled for near real-time visibility
Trade status tracking: active to paused to stopped to completed
Performance viewer with 40+ metrics - Sharpe, Sortino, Calmar, SQN, win ratio, drawdown
Request management
Backtest, livetest, livetrade and criteria requests from the dashboard.
Live trade tracking
Positions and orders with P&L, entry/close prices and durations.
Strategy browser
Browse Strategy Information Capsules and their configuration.
Engine health
Heartbeat and log monitoring for the trading engine.
| Metric | Value |
|---|---|
| MongoDB collections | 0+ |
| performance metrics | 0+ |
| markets covered | 0 |
| frontend developer | Solo |
I built the entire operator dashboard - the human window into an autonomous trading engine. That meant modeling 12+ MongoDB collections into a coherent UI: request flows, the live-trade lifecycle, positions and orders monitoring, a 40+ metric performance viewer, a strategy browser, and engine heartbeat/log health.
- Visualizing an autonomous system is a data-modeling problem first and a UI problem second.
- Operators trust a control room only when state is consistent and the lifecycle is legible.
